arXiv:2607.04206v1 Announce Type: new Abstract: Diffusion language models (dLLMs) generate text by iteratively denoising a masked response and can commit multiple output positions per model invocation. Their bidirectional attention prevents exact autoregressive-style KV caching, since committing one position shifts the KV activations of all others. Approximate caching techniques such as Fast-dLLM and dKV-Cache refresh KV activations repeatedly and reuse them across intervening decodes, inducing a repeated prefill/decode structure. This makes AR serving mechanisms relevant to dLLMs, but not directly applicable. dLLM decodes are block-sized rather than token-sized, prefills recur, and bidirectional attention precludes the chunked prefill mechanism used for stall-free colocated serving. We present Sangam, a serving system for cached dLLM inference. Sangam introduces a deficit token-budget scheduler that admits in-flight decodes first, admits whole indivisible prefills only when the accumulated token budget allows, and carries unused budget forward. This achieves amortized stall-free scheduling. Disaggregated serving avoids prefill-decode interference but suffers from prefill/decode resource partitioning problem. Sangam adopts a hybrid serving strategy, overflowing prefills onto decode workers to relieve prefill under-provisioning, and uses the same deficit-budget scheduler to protect those workers' decodes from the overflow. We show that like AR serving, dLLM serving design space is governed by prefill-decode interference and prefill/decode partitioning. Colocated serving is most effective on decode-heavy workloads, cutting mean latency by 9-20% over hybrid execution on LLaDA-8B ShareGPT; while hybrid execution is most effective on prefill-heavy workloads, cutting mean latency by 8-20% over colocated execution on Dream-7B arXiv. Sangam is available at https://github.com/UT-InfraAI/sangam.
Science Journals
arXiv:2607.04212v1 Announce Type: new Abstract: Role-based multiagent code generation aims to make LLMs more effective on repository-scale problems, moving beyond small programming tasks. We evaluate this approach on 12 Java repositories, finding greater similarity to developer code than single LLMs, but a persistent gap from human implementations.
arXiv:2607.04218v1 Announce Type: new Abstract: The growing demand for privacy-preserving, data-intensive applications such as IoT, augmented reality, and autonomous systems positions Federated Learning (FL) as a key enabler in 6G networks. Over-the-Air FL (OTA-FL) leverages the superposition property of the wireless multiple access channel for efficient aggregation via simultaneous transmissions. Existing methods rely on fixed aggregation schedules and do not jointly address noise, fading, and client heterogeneity. We propose CHARGE-FL (CHannel-Adaptive Robust agGrEgation), a framework that adaptively schedules aggregation based on channel dynamics and application readiness. By combining a tailored optimization strategy with a dual-purpose precoding mechanism, CHARGE-FL mitigates channel distortion and bias from partial updates, achieving superior accuracy, stability, and convergence under realistic wireless conditions. Empirical results under realistic wireless conditions show that CHARGE-FL significantly improves accuracy, stability, and convergence over state-of-the-art OTA-FL methods, particularly in straggler-prone and noisy scenarios.
arXiv:2607.04234v1 Announce Type: new Abstract: Deformable object manipulation poses challenges beyond task completion: successful execution must also maintain safe physical interaction, holding the object stably without slip or drop while avoiding excessive deformation. However, existing manipulation benchmarks are predominantly success-oriented and rarely evaluate whether a policy remains physically safe throughout execution. We present SoftVTBench, a safety-aware visuo-tactile benchmark for physically constrained deformable object manipulation. Built in Isaac Sim with finite-element-simulated deformable objects, SoftVTBench provides multi-view RGB observations, RGB tactile sensing with marker motion, proprioception, and language instructions, and defines four matched task suites over object type (deformable vs. rigid) and variation axis (object vs. spatial). It separately reports Goal Success and Safety Success; the latter additionally requires no drop and peak deformation below a calibrated object-specific threshold, measured from policy-hidden privileged Finite Element Method (FEM) states. We implement pi0.5-based baselines under this protocol. Experiments show that success-only evaluation substantially overstates policy performance, as a large fraction of goal-completing rollouts still violate physical safety. Furthermore, incorporating tactile sensing improves Safety Success (e.g., from 21.4% to 35.6% on object-centric deformable tasks) and reduces object deformation during execution, while maintaining comparable Goal Success. SoftVTBench provides a reproducible benchmark for studying visuo-tactile deformable manipulation under physical interaction constraints.
arXiv:2607.04944v1 Announce Type: new Abstract: Wavelength-shifting (WLS) techniques enable particle detectors based on noble gases, whose scintillation light is predominantly emitted in the vacuum-ultraviolet. We investigate WLS fibers coated with tetraphenyl butadiene (TPB) for scintillation light detection in gaseous xenon and argon at pressures up to 8.5 bar, motivated by future high-pressure xenon time-projection chambers of the NEXT program. Two detector configurations are studied: an elongated high-pressure vessel with four PTFE panels equipped with WLS fibers read by temperature-stabilized SiPMs, and a compact box-shaped detector operated at 1 bar Xe with WLS fibers read out by PMTs. Both operate with continuous gas purification. The detector response is characterized using cosmic muons and alpha particles from a $^{241}$Am source. With the SiPM setup, we measure a light collection efficiency (LCE) of ${1.18 \pm 0.01~\mathrm{(sta.)}~^{+0.07}_{-0.09}~\mathrm{(sys.)}~\%}$ for xenon and ${1.07 \pm 0.01~\mathrm{(sta.)}~^{+0.06}_{-0.08}~\mathrm{(sys.)}~\%}$ for argon. With PMT readout, we measure a LCE of ${0.45 \pm 0.01~\mathrm{(sta.)} \pm 0.05~\mathrm{(sys.)}~\%}$ in xenon, in agreement with the SiPM result once photon detection efficiency is accounted for. Average scintillation waveforms in xenon and argon are studied to assess the time structure of the emitted light. Cosmic-muon measurements yield a mean energy required to produce a scintillation photon $45\pm7~\mathrm{(sta.)}~^{+4}_{-5}~\mathrm{(sys.)}~\mathrm{eV}$ at 1.5 bar, in agreement with the literature. The results demonstrate that TPB-coated WLS fiber systems can reliably detect scintillation light in high-pressure gaseous noble detectors, with a LCE representing an upper limit for realistic large-scale TPCs, where additional photon losses from materials and fiber attenuation are expected.
arXiv:2607.04963v1 Announce Type: new Abstract: Reinforcement Learning (RL) is the dominant paradigm for training Large Language Model (LLM) agents on long-horizon tasks. However, sparse and delayed rewards often lead to trajectory neglect, in which agents lose focus on the task goal and interaction history at intermediate steps. Prior work has explored step-level supervision using Shannon-entropy-based uncertainty signals, which conflate inherent state complexity with agent confidence and therefore provide unreliable estimates of decision reliability. To address this issue, we propose normalized entropy, which measures confidence deviations relative to an agent's average behavior under a given state, thereby strengthening the association between low-quality actions and trajectory neglect. Building on this insight, we introduce Selective Trajectory-Aware Policy Optimization (STAPO), a hierarchical group-based RL framework. STAPO leverages normalized entropy to locate outlier steps associated with trajectory neglect and optimizes them via a joint mechanism of trajectory-aware reward and trajectory-independent penalty, enhancing trajectory awareness while preserving training stability. Extensive experiments on ALFWorld, WebShop, and Search-Augmented QA demonstrate that STAPO achieves state-of-the-art performance while substantially alleviating trajectory neglect, validating its effectiveness and robustness for agentic tasks.
arXiv:2607.04945v1 Announce Type: new Abstract: LLMs enable the integration of external conceptual resources at inference time, creating new opportunities for detecting ideologically and historically complex phenomena such as antisemitism. We investigate how different forms of conceptual grounding affect antisemitism detection and explanation behavior across four state-of-the-art LLMs. Using two expert-annotated datasets, we compare definitional, fine-grained taxonomic, example-augmented, and large-context representations of antisemitism. We find that fine-grained taxonomic representations substantially improve recall, while simultaneously reducing precision. Surprisingly, supplying substantially larger conceptual resources yields no additional quantitative benefit. Post-Holocaust antisemitism poses the most persistent challenge across models and configurations. Analysis of explanations further reveals systematic limitations including overproduction of conceptual references, reliance on lexical cues, overconfidence, and difficulties with subtle or justificatory forms of antisemitism. Our findings highlight both the potential and the remaining limitations of conceptually grounded LLMs for antisemitism detection and reasoning.
arXiv:2607.04566v1 Announce Type: new Abstract: Adolescence is a critical developmental period marked by heightened emotional sensitivity, social stress, and vulnerability to substance use. However, traditional research methods provide limited access to adolescents' authentic experiences, hindering efforts to develop evidence-based prevention and intervention strategies. Social media provides a unique opportunity to observe adolescents' naturally occurring discussions about substance use, offering valuable insights into their opinions, emotions, and lived experiences that can inform early prevention and intervention strategies. In this study, we analyze large-scale Reddit discussions related to substance use among adolescents between 2018 and 2023. Leveraging hour-by-day temporal analysis, sentiment and emotion classification, and transformer-based topic modeling (BERTopic), we examine the interaction between time, emotion, and semantic content in adolescent substance use discourse. Our findings reveal pronounced weekend and late-night peaks in substance-related discussions, a dominance of negative emotions such as sadness and fear, and distinct semantic topics centered on peer relationships, family conflict, emotional distress, and substance-specific experiences. These findings advance our understanding of adolescent substance use in naturalistic online settings and provide empirical evidence to support the development of more timely, targeted, and evidence-based prevention and intervention strategies.
arXiv:2607.04425v1 Announce Type: new Abstract: Recent advances in multimodal foundation models and agent systems have driven GUI agents from single-platform task execution toward cross-platform interaction. However, building multi-platform GUI agents remains challenging. On one hand, high-quality and executable cross-platform interaction trajectories are still scarce, and existing data often suffer from limited platform coverage. On the other hand, different platforms exhibit distinct interaction conventions, making joint or continual training prone to behavioral pattern mixing, platform-specific capability degradation, and catastrophic forgetting. To address these challenges, we construct Uni-GUI, a high-quality cross-platform GUI interaction dataset, and propose UI-MOPD, the first method that incorporates multi-teacher on-policy distillation into continual learning for GUI agents. UI-MOPD dynamically selects a platform-specific teacher according to the current environment and transfers platform-specific behavioral priors to a shared policy through platform-conditioned distillation, enabling adaptation to new platforms while preserving capabilities on existing ones. Experiments on OSWorld and MobileWorld show that UI-MOPD achieves task success rates of 38.2% and 12.0%, respectively, demonstrating its effectiveness in balancing cross-platform capability retention and new-platform adaptation. Project page: https://elispectre.github.io/UI-MOPD/.
arXiv:2607.04430v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in question answering (QA) systems, yet they may generate hallucinated or misaligned responses without reliable confidence estimates. Uncertainty quantification (UQ) offers a natural basis for selective answering, where a system answers only when its prediction is deemed reliable and abstains otherwise. However, existing uncertainty scores for LLMs are often heuristic: a threshold chosen on such scores does not, by itself, provide statistical guarantees on the error rate among accepted answers. We propose CIC, a confidence-interval-based calibration framework that converts arbitrary uncertainty scores into risk-controlled selective answering rules. Given a held-out calibration set, CIC evaluates each generated response using an application-specific alignment criterion and associates it with an uncertainty score and a binary error label. For each candidate uncertainty threshold, CIC estimates the acceptance-conditioned error rate and constructs a high-probability upper confidence bound using either Hoeffding-style or Clopper-Pearson confidence intervals. It then selects the largest threshold whose upper bound is below a user-specified risk level $\alpha$, thereby maximizing the answering rate subject to a finite-sample reliability constraint. Under exchangeability, CIC guarantees with probability at least $1-\delta$ that the selected threshold, if non-null, controls the error rate among accepted answers at level $\alpha$. We evaluate CIC on both closed-ended and open-ended QA benchmarks across seven LLMs and multiple uncertainty estimators. Experimental results show that CIC consistently achieves valid risk control while retaining strong answering efficiency, providing a practical and statistically grounded mechanism for deploying LLMs in reliability-sensitive QA workflows.
arXiv:2607.04447v1 Announce Type: new Abstract: Local causal discovery is a scalable alternative to global structure learning. However, it can struggle to identify valid adjustment sets in data-scarce settings because of finite-sample uncertainty, incomplete local neighborhoods, and unresolved Markov equivalence. Although many application domains provide structured background knowledge, its integration into local causal discovery remains limited. We propose b-LOAD, a knowledge-informed extension of the LOAD algorithm for local discovery of optimal adjustment sets. b-LOAD incorporates prior edge constraints directly into the local structure-learning procedure and uses Meek's rules to expand the discovery frontier dynamically, yielding a knowledge-constrained partially directed graph over the relevant local subgraph. This strategy helps prevent structurally relevant nodes introduced by prior knowledge from being excluded by local search. We prove that, under sound background knowledge, the procedure monotonically refines the admissible equivalence class and can enlarge the set of identifiable causal queries, enabling recovery of optimal adjustment sets that are not identifiable from observational conditional-independence information alone. Empirically, b-LOAD improves downstream causal effect estimation relative to purely data-driven and standard knowledge-augmented baselines, particularly in data-scarce and structurally complex regimes. Results on real-world biological networks show that locally targeted prior knowledge provides the largest gains and remains beneficial under moderate structural noise. These findings position b-LOAD as a scalable approach for converting fragmented domain knowledge into more reliable causal-effect estimation.
arXiv:2607.04428v1 Announce Type: new Abstract: Diffusion large language models (dLLMs) generate text by iteratively denoising a masked sequence, offering a parallel alternative to autoregressive models, but eliciting strong reasoning through post-training remains difficult: supervised fine-tuning is off-policy and suffers from exposure bias, while reinforcement learning gives only sparse, sequence-level rewards and is hard to apply without tractable sequence likelihoods. On-policy self-distillation (OPSD) offers a promising alternative, using one model as both student and teacher to provide dense, token-level, on-policy supervision, but its effectiveness hinges on giving the teacher privileged information (PI) - typically an instance-specific ground-truth reference unavailable at inference - so the student ends up distilling a weak PI-free consensus policy that yields little improvement on dLLM reasoning. We introduce dOPSD, which instead derives the teacher's privilege directly from the student's own denoising trajectory, evaluating masked positions using later, more-decoded steps of that same trajectory rather than an external label, so the teacher's advantage emerges from the model's own decoding process; on Dream and LLaDA, dOPSD improves both in-domain math reasoning and out-of-domain code generation, outperforming supervised and on-policy baselines.
arXiv:2607.04967v1 Announce Type: new Abstract: We present a method for estimating the hydrodynamic coefficients in the Cummins equations using time-series data from a moving body, such as a floating offshore structure. The proposed data-driven method is based on incorporating the Cummins equations governing the dynamics of a structural body interacting with water waves into a physics-informed neural network (PINN), along with available motion data. The proposed method first estimates the structure's state in terms of translational and rotational degrees of freedom, and then solves the inverse problem to determine the hydrodynamic forces acting on the body, expressed in terms of added mass, damping coefficients, and/or hydrostatic restoring. The Cummins equations are formulated as a first-order system, and both state and parameter estimation are performed using PINNs. The method is verified on the free decay of a sphere and a box. The results demonstrate that it is possible to estimate the state and hydrodynamic coefficients accurately, although accuracy depends on the volume and quality of the movement data.
arXiv:2607.04503v1 Announce Type: new Abstract: This article examines the institutional conditions under which artificial intelligence systems in U.S. welfare administration come to operate as instruments of support or as instruments of control. Rather than asking what welfare algorithms "really" are (tools of proactive assistance or infrastructures of surveillance) the article starts from the premise that support and control are co-present within the same system, while their relative balance shifts over time. This movement is conceptualized through the notion of support-control convergence and the model of an institutional ratchet. Routine budgetary and political pressures make control-oriented effects easily measurable and politically capitalizable, whereas a return toward support requires external intervention of disproportionate force, such as judicial compulsion, legislative prohibition, or public scandal. Empirically, the article draws on process tracing of six state- and county-level cases: NYSDOL fraud detection, Michigan MiDAS, Illinois Medicaid managed care, LA County homelessness prevention, the Allegheny Family Screening Tool, and Washington Foster Care. The findings show that the system's orientation is shaped by institutional design, with the decisive parameter being the side on which the costs of algorithmic error are placed. Drift toward control is routine, while reversal is exceptional and costly. In the MiDAS case, activation required a single administrative decision, whereas reversal took nine years and a $20 million settlement; even then, the system did not return to a support-oriented configuration.
arXiv:2607.04511v1 Announce Type: new Abstract: Enterprise security teams commonly summarize remediation through mean time to remediate (MTTR), SLA compliance, dwell time, or detection delay. These metrics are useful, but they can hide how fixes actually reach the estate: continuously, through scheduled maintenance windows, in deployment rings, or through emergency bypass paths. This paper introduces a remediation-cadence audit for enterprise vulnerability management. The audit records routine mean lag, release period, release fraction, cohort geometry, emergency/routine split, non-fielding delay, local residual-pressure evidence, and declared rate scenario. It compares a continuous same-mean shortcut with the recorded release calendar and reports a local capacity verdict plus a calendar discount: the fraction of mean-only local capacity consumed by calendarized fielding. Worked notional packets with the same 30-day mean lag show why this matters. Under the normalized screening scenario, a two-month release train consumes 17.4\% of mean-only capacity, a monthly train 5.2\%, and a two-week screen 1.3\%; across a 16-fold attacker-adjustment rate band, the two-month discount remains at least about 12\% and the monthly discount stays in the resolution-sensitive 3--8\% range. The audit therefore turns cadence assessment into an evidence-resolution question: when the discount is material relative to residual-pressure uncertainty or claimed headroom, MTTR/SLA should not be used alone as fielding evidence. Release-geometry checks show that deployment rings do not automatically recover the continuous benchmark, and cohort staggering can help or hurt near capacity. The result is a reproducible governance diagnostic, not a breach predictor or CVE prioritizer.
arXiv:2607.04525v1 Announce Type: new Abstract: How concepts are represented in neural networks is a fundamental question in machine learning. The dominant view treats concept representations as stationary geometric objects. Yet concepts appear in context, and context transforms them. Drawing from neural population geometry, we formalize concept representations as point-cloud manifolds and contextual transformations as vector fields, and instantiate this framework in large language models. Across six model families of varying scales, we find that context moves each concept differently. The variance in these displacements is semantically organized, correlating with lexical concreteness and density. Importantly, both the concepts being transformed and this variance structure are shared across models: displacement structure transported from one model predicts held-out displacements in others significantly above chance. Together, these findings show that models share a common geometry not only in how concepts are represented, but more importantly in how context transforms them, a structure with richer organization than prior work has recognized.
arXiv:2607.04982v1 Announce Type: new Abstract: High-resolution velocity models are crucial for reservoir characterization and subsurface delineation. However, the band limited nature of our surface recorded data limits resolution. Utilizing well measurements to enhance the resolution of our subsurface models is an important objective. To this end, we present a diffusion-guided framework for structurally preconditioned velocity-model reconstruction from sparse well-log information. The proposed approach combines plane-wave PDE regularization, structurally preconditioned inversion, and measurement-guided diffusion posterior sampling within a unified formulation. Local structural slopes estimated through plane-wave destruction are used both to propagate well information along geological dip directions and to guide the diffusion sampling process through a joint velocity--slope generative prior. Numerical experiments on the Volve synthetic model and the Viking Graben field dataset demonstrate that the proposed framework improves structural continuity, lateral consistency, and geological realism compared with conventional structurally preconditioned inversion approaches while maintaining computationally practical inference through DDIM sampling.
arXiv:2607.04557v1 Announce Type: new Abstract: Accurate prediction of patient-specific therapeutic response from pre-treatment transcriptomes is hindered by the scarcity of matched clinical response labels and post-treatment molecular profiles. Preclinical transfer-learning models can simulate drug-induced expression changes but are often hard to interpret and unstable, whereas knowledge-graph methods provide mechanistic context yet remain static and fail to capture drug-induced transcriptomic perturbation dynamics. We propose PREDIKTOR, a patient-centered multi-view framework that aligns a personalized network view with a transferable transcriptomic perturbation view to predict clinical drug response. For each patient, we construct an individualized gene regulatory network from tumor expression using DysRegNet and augment it with drug-target links from DrugBank; a graph neural encoder yields a drug-centric, mechanistically grounded embedding. In parallel, a frozen condition-specific gene-gene attention model pretrained on LINCS L1000 generates a simulated post-perturbation transcriptomic profile for the same patient-drug pair. We align the two views in a shared latent space via a CLIP-style contrastive objective with drug-context hard negatives, then concatenate the representations for end-to-end response classification. On TCGA, PREDIKTOR consistently outperforms state-of-the-art baselines under patient-, drug-, and tissue-split evaluations, and transfers zero-shot to the I-SPY2 trial, improving AUROC by 5.6% over competing methods. The aligned embeddings yield stable gene and pathway attributions that recover known mechanisms, supporting actionable and interpretable precision oncology.
arXiv:2607.04516v1 Announce Type: new Abstract: This study presents a near-real-time, meter-scale three-dimensional urban wind simulation framework for low-altitude flight events in complex urban meteorological environments. It reconstructs high-resolution wind fields by combining sparse observations with efficient microscale flow modeling. The framework integrates lattice Boltzmann method large-eddy simulation (LBM-LES), high-fidelity urban morphology reconstruction that explicitly resolves real building details, and observation-driven boundary assimilation into a rapid end-to-end pipeline for realistic urban domains. Multi-site Doppler lidar measurements from dense urban Guangzhou, China, are used for evaluation. The system reconstructs three-dimensional wind fields at 5 m resolution over kilometer-scale domains within minutes. Robustness and accuracy are tested through controlled observation reduction, independent validation against withheld lidar stations, and sensitivity analyses of grid resolution and precursor domain extent. Results show stable reproduction of vertical wind structures and key local flow features under complex morphology and limited observations, providing a scalable pathway for near-real-time urban wind reconstruction.
arXiv:2607.04542v1 Announce Type: new Abstract: Every LLM agent run re-derives its behavior token by token on a frontier model: brilliant, expensive, slow, and unbounded. We present Auto, a compiler that records live agent behavior, measures which parts are secretly deterministic, extracts them into verified programs or distilled specialists, and emits cognition binaries: WebAssembly artifacts whose manifests carry measured guarantees and whose declared capabilities are physically enforced by the sandbox. A tiered runtime executes compiled behavior behind conformally calibrated guards; guard trips deopt to the reference agent, and the captured trace recompiles back down, so nothing is figured out twice. We use "AGI compiler" in one narrow, testable sense: a system that autonomously converts novel experience into permanent, verified, near-free skill while measuring what it does not know. On AUTO-BENCH, a benchmark we introduce and pre-register, 87.1% of 560 recorded frontier-agent spans are witnessed-deterministic (three of the four censused task families measure 100.0%). On a 300-item stream with three scheduled distribution shifts, the closed loop compiles three artifact generations and drives marginal cost from 59 to 2 micro-dollars per item (6.4x end-to-end) at 96.9% parity on witnessed inputs with zero errors. The same stream also quantifies the failure modes: a loose guard silently mislabels 48.9% of compiled answers, and an unfaithful deopt reference causes the verification gate to refuse recompilation. Calibration and reference fidelity, not model capability, decide whether cheap stays correct. Code: https://github.com/RightNow-AI/auto
arXiv:2607.04543v1 Announce Type: new Abstract: Governments are important actors in frontier AI governance, but many facts about their adoption and use of AI systems are difficult to observe directly. Procurement disclosures and official statements are useful, but can also be delayed, selective, and better suited to measuring formal adoption than actual day-to-day use. We propose a complementary monitoring primitive: measuring traces of language-model assistance in public government documents. The approach is lightweight, externally reproducible, and based on revealed behavior rather than stated intent. In a pilot study of ten public document streams from U.S. and PRC government-related sources, we find that, while 2021 baselines are consistently near zero, by 2026, four of our ten sources show statistically significant signs of AI-assisted writing. In our sample, the U.S. signal concentrates in publications downstream of policy work; the PRC signal concentrates closer to it. We close by discussing how this signal could complement existing instruments for monitoring government AI adoption, and where it falls short.
arXiv:2607.04509v1 Announce Type: new Abstract: We consider the 2D RMQ encoding problem: given an $m\times n$ array of $mn$ elements over a total order, encode it such that, for any query rectangle, the position of its maximum element can be reported without accessing the original array. For $m \le n$, it is known how to encode the array in $O(mn \min\{m, \log n\})$ bits with $O(1)$-time queries [Brodal et al., Algorithmica 2012], and also how to obtain an asymptotically optimal encoding consisting of $O(mn \log m)$ bits [Brodal et al., ESA 2013]. However, the latter approach does not prove any guarantee on the query time, and it appears to be inherently sequential: it requires scanning the whole encoding to answer a query. We design a different encoding that uses near-optimal space while allowing for efficient queries. More concretely, for every parameter $\kappa\in[1, \log\log n]$, our encoding uses $O(\kappa mn(\log m+\log\log n))$ bits and answers 2D RMQ queries in $O(\log^{1/\kappa}n)$ time.
arXiv:2607.04547v1 Announce Type: new Abstract: People respond to artificial intelligence chatbots (AICs) in highly variable ways. In this paper, we adapt Bronfenbrenner's theory into a heuristic framework for understanding this variation. The framework places the human user at the center while also placing the AI there and reconceptualizing the proximal processes as the repeated, reciprocal, and coadaptive interactions between the user and a personalized AIC. The surrounding systems identify the contextual factors that shape how the user experiences, interprets, responds to, and is changed by these interactions. Because stateful AICs learn from accumulated exchanges with their users and have memory, users are responding not only to an AIC but also to a version of the AIC that their own prior interactions have helped create. This extension preserves Bronfenbrenner's emphasis on proximal processes while accounting for the unique dynamics of personalized AICs. The resulting framework provides a structured map of where and how variation in human and AIC relationships arises, as well as having implications for researchers, practitioners, and AIC designers.
arXiv:2607.04167v1 Announce Type: new Abstract: Recent work showed that language models represent character counts on curved 1D manifolds, with attention heads performing geometric transformations to enable computation. We test whether this generalizes across four ordinal tasks (bracket depth, indentation, table position, numeric magnitude) in Gemma-2-2B, Gemma-2-9B, and Qwen3-4B. We find that 1D manifolds with place-cell feature tiling emerge for tasks where the ordinal variable is locally computable from token identity, while tasks requiring cross-position integration or semantic extraction produce higher-dimensional or incoherent representations. Geometric computation is architecture-dependent: Qwen3-4B shows substantially stronger twisting than Gemma models for indentation, and its twisters preserve ordinal order, unlike its numeric twisters. Activation patching confirms that the identified manifold subspaces concentrate task-relevant information, with manifold-direction ablation causing dramatically larger probe accuracy drops than random-direction controls.
arXiv:2607.04184v1 Announce Type: new Abstract: Market manipulation is the dubious practice of manipulating stock prices in order to make a quick profit, which truly degrades confidence on trading platforms. We implemented an unsupervised fraud-detection toolkit that begins with K-Means++ clustering to address this issue. A dataset of roughly one million financial transactions from 2012 to 2024 is used. In order to identify fraudulent trades and categorize them using market practice heuristic thresholds, the study suggests a clustering-based pipeline. The method highlights 2.02% of trades as suspicious where 51.10% clearly indicate spoofing, 0.10% indicate pump and dump, 0.55% indicate insider trading, 1.43% indicate a fake breakout, and 46.83% are unclassified. Despite the lack of ground truth, the model's performance is confirmed by a Silhouette Score of 0.561.